One in 20 students was a victim of violence or theft at school in 2003, the government said in a report that shows school crime rates are about half what they were 10 years earlier.
Nevertheless, the school crime rate essentially has leveled off, showing no significant change since 2000, according to a report yesterday from the Education and Justice departments.
There were about 28 incidents of rape, sexual assault, robbery and physical assault for every 1,000 students in 2003, compared with 59 per 1,000 a decade earlier. The study looked at crimes against the 26.4 million students who were 12 to 18 in 2003.
In 2002, the violent-crime rate per 1,000 students was 24, but government researchers said there was no statistically significant change from 2002 to 2003 because the numbers are estimates from relatively small surveys.
“The level of precision isn’t good enough to say whether there has been a change,” said Thomas Snyder, a report author at the Education Department.
Mr. Snyder, however, said there has been no change in the crime rate in several years. The report does not attempt to explain rises and falls.
In 2003, there were about 738,700 violent crimes involving students at school and about 846,400 away from school property. For the most serious nonfatal violent crimes — rape, assault and robbery — the rates were at least 50 percent lower in school than away from school every year from 1992 to 2003.
Students were twice as likely to be victims of serious violent crimes away from school than at school, but more likely to have things stolen from them at school than elsewhere.
Students from poorer families were more likely to be victims of violent crimes at school than were wealthier students, while the opposite was true for theft, with richer students more likely to be victims.
Some school safety specialists have attributed the fall in the crime rate in the past decade to more metal detectors, security personnel and programs aimed at curbing bullying, which can lead to more serious crimes.
But some critics said the annual report routinely understates crime in schools because it is based on limited surveys and self-reporting. The data also are outdated, said Kenneth Trump, president of National School Safety and Security Services, a consulting firm in Cleveland.
“Even if the government had actual real-time data, it’s two to three years old,” Mr. Trump said. “School administrators need to know what is happening today and what to anticipate tomorrow, not outdated numbers from yesteryear.”